Skip to main content
. 2010 Nov 15;11:562. doi: 10.1186/1471-2105-11-562

Table 3.

Improvement in F1 score for H. sapiens PPI dataset

Best-match average Maximun
IEA- IEA+ IEA- IEA+
CC BP MF CC BP MF CC BP MF CC BP MF
3.44 1.51 2.42 1.28 1.64 4.0 2.7 1.48 2.0 1.53 1.58 1.50

Average improvement in F1 scores achieved by TCSS over Resnik for best-match average and maximum approaches. TCSS does 2.2 times better than Resnik for cellular component (CC), 1.5 times for biological process (BP), and 2.5 times for molecular function (MF) ontologies on average.